drm/i915: Kick hangcheck from retire worker
authorChris Wilson <chris@chris-wilson.co.uk>
Sat, 9 Jul 2016 09:12:06 +0000 (10:12 +0100)
committerChris Wilson <chris@chris-wilson.co.uk>
Mon, 11 Jul 2016 12:48:42 +0000 (13:48 +0100)
Let's ensure that we cannot run indefinitely without the hangcheck
worker being queued. We removed it from being kicked on every request
because we were kicking it a few millions times in every hangcheck
interval and only once is necessary! However, that leaves us with the
issue of what if userspace never waits for a request, or runs out of
resources, what if userspace just issues a request then spins on
BUSY_IOCTL?

Testcase: igt/gem_busy
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1468055535-19740-3-git-send-email-chris@chris-wilson.co.uk
Reviewed-by: Mika Kuoppala <mika.kuoppala@intel.com>
drivers/gpu/drm/i915/i915_gem.c

index 7fd44980798fd048714e927e8a305b94795175a0..adeca0ec4cfb3b37c67f6f194016c66d7eeddb65 100644 (file)
@@ -3281,10 +3281,12 @@ i915_gem_retire_work_handler(struct work_struct *work)
         * We do not need to do this test under locking as in the worst-case
         * we queue the retire worker once too often.
         */
-       if (READ_ONCE(dev_priv->gt.awake))
+       if (READ_ONCE(dev_priv->gt.awake)) {
+               i915_queue_hangcheck(dev_priv);
                queue_delayed_work(dev_priv->wq,
                                   &dev_priv->gt.retire_work,
                                   round_jiffies_up_relative(HZ));
+       }
 }
 
 static void